108 \end{center} |
179 \end{center} |
109 |
180 |
110 \normalsize |
181 \normalsize |
111 \begin{center} |
182 \begin{center} |
112 \begin{tabular}{ll} |
183 \begin{tabular}{ll} |
113 Email: & christian.urban at kcl.ac.uk\\ |
184 %Email: & christian.urban at kcl.ac.uk\\ |
114 Office: & N\liningnums{7.07} (North Wing, Bush House)\\ |
185 %Office: & N\liningnums{7.07} (North Wing, Bush House)\\ |
115 Slides: & KEATS |
186 %Slides: & KEATS |
|
187 Student: & Chengsong Tan\\ |
|
188 Supervisor: & Christian Urban \\ |
|
189 Date: & 2019/5/10 |
116 \end{tabular} |
190 \end{tabular} |
117 \end{center} |
191 \end{center} |
118 |
192 |
119 \end{frame} |
193 \end{frame} |
120 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
194 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
121 |
195 |
122 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
123 \begin{frame}[t] |
|
124 \frametitle{Why Study Compilers?} |
|
125 |
|
126 John Regehr {\small(Univ.~Utah, LLVM compiler hacker)}\smallskip\\ |
|
127 |
|
128 \begin{bubble}[10.5cm] |
|
129 \bf ``\ldots{}It’s effectively a perpetual |
|
130 employment act for solid compiler hackers.'' |
|
131 \end{bubble} |
|
132 |
|
133 \onslide<1->{ |
|
134 \only<2>{ |
|
135 \begin{itemize} |
|
136 \item {\bf Hardware is getting weirder |
|
137 rather than getting clocked faster} |
|
138 |
|
139 \begin{itemize} |
|
140 \item Almost all processors are |
|
141 multicores nowadays and it looks like there is increasing asymmetry in |
|
142 resources across cores. Processors come with vector units, crypto |
|
143 accelerators etc. We have DSPs, GPUs, |
|
144 ARM big.little, and Xeon Phi. This is only scratching the |
|
145 surface. |
|
146 \end{itemize} |
|
147 \end{itemize}} |
|
148 \only<3>{ |
|
149 \begin{itemize} |
|
150 \item {\bf We’re getting tired of low-level languages and |
|
151 their associated security disasters} |
|
152 |
|
153 \begin{itemize} |
|
154 \item |
|
155 We want to write new code, to |
|
156 whatever extent possible, in safer, higher-level |
|
157 languages. Compilers are caught right in the middle of these |
|
158 opposing trends: one of their main jobs is to help bridge the large |
|
159 and growing gap between increasingly high-level languages and |
|
160 increasingly wacky platforms. |
|
161 \end{itemize} |
|
162 \end{itemize}}} |
|
163 |
|
164 \end{frame} |
|
165 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
166 |
|
167 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
168 \begin{frame}[c] |
|
169 \frametitle{Why Bother?} |
|
170 |
|
171 \begin{columns}[t] |
|
172 \begin{column}{.5\textwidth} |
|
173 Ruby, Python, Java 8\medskip\\ |
|
174 \begin{tikzpicture}\footnotesize |
|
175 \begin{axis}[ |
|
176 xlabel={$n$}, |
|
177 x label style={at={(1.05,0.0)}}, |
|
178 ylabel={time in secs}, |
|
179 enlargelimits=false, |
|
180 xtick={0,5,...,30}, |
|
181 xmax=33, |
|
182 ymax=35, |
|
183 ytick={0,5,...,30}, |
|
184 scaled ticks=false, |
|
185 axis lines=left, |
|
186 width=5.5cm, |
|
187 height=4cm, |
|
188 legend entries={Python,Ruby}, |
|
189 legend pos=north west, |
|
190 legend cell align=left] |
|
191 \addplot[blue,mark=*, mark options={fill=white}] table {re-python.data}; |
|
192 \addplot[brown,mark=triangle*, mark options={fill=white}] table {re-ruby.data}; |
|
193 \end{axis} |
|
194 \end{tikzpicture} |
|
195 \begin{tikzpicture}\footnotesize |
|
196 \begin{axis}[ |
|
197 xlabel={$n$}, |
|
198 x label style={at={(1.05,0.0)}}, |
|
199 ylabel={time in secs}, |
|
200 enlargelimits=false, |
|
201 xtick={0,5,...,30}, |
|
202 xmax=33, |
|
203 ymax=35, |
|
204 ytick={0,5,...,30}, |
|
205 scaled ticks=false, |
|
206 axis lines=left, |
|
207 width=5.5cm, |
|
208 height=4cm, |
|
209 legend entries={Python, Java 8}, |
|
210 legend pos=north west, |
|
211 legend cell align=left] |
|
212 \addplot[blue,mark=*, mark options={fill=white}] table {re-python2.data}; |
|
213 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java.data}; |
|
214 \end{axis} |
|
215 \end{tikzpicture} |
|
216 |
|
217 \end{column} |
|
218 \begin{column}{.5\textwidth} |
|
219 Us (after next lecture)\medskip\\ |
|
220 \begin{tikzpicture}\footnotesize |
|
221 \begin{axis}[ |
|
222 xlabel={$n$}, |
|
223 x label style={at={(1.07,0.0)}}, |
|
224 ylabel={time in secs}, |
|
225 enlargelimits=false, |
|
226 xtick={0,5000,...,10000}, |
|
227 xmax=11000, |
|
228 ymax=35, |
|
229 ytick={0,5,...,30}, |
|
230 scaled ticks=false, |
|
231 axis lines=left, |
|
232 width=5.5cm, |
|
233 height=4cm] |
|
234 \addplot[green,mark=square*,mark options={fill=white}] table {re2.data}; |
|
235 \addplot[black,mark=square*,mark options={fill=white}] table {re3.data}; |
|
236 \end{axis} |
|
237 \end{tikzpicture} |
|
238 \begin{tikzpicture}\footnotesize |
|
239 \begin{axis}[ |
|
240 xlabel={$n$}, |
|
241 x label style={at={(1.07,0.0)}}, |
|
242 ylabel={time in secs}, |
|
243 enlargelimits=false, |
|
244 ymax=35, |
|
245 ytick={0,5,...,30}, |
|
246 scaled ticks=false, |
|
247 axis lines=left, |
|
248 width=5.5cm, |
|
249 height=4cm] |
|
250 \addplot[black,mark=square*,mark options={fill=white}] table {re3a.data}; |
|
251 \end{axis} |
|
252 \end{tikzpicture} |
|
253 \end{column} |
|
254 \end{columns}\bigskip |
|
255 |
|
256 \small\centering |
|
257 matching \texttt{[a?]\{n\}[a]\{n\}} and \texttt{(a*)*b} |
|
258 against $\underbrace{\texttt{a}...\texttt{a}}_n$ |
|
259 \end{frame} |
|
260 |
|
261 \begin{frame} |
|
262 \begin{column}{.5\textwidth} |
|
263 \begin{tikzpicture} |
|
264 \begin{axis}[ |
|
265 xlabel={$n$}, |
|
266 x label style={at={(1.05,0.0)}}, |
|
267 ylabel={time in secs}, |
|
268 y label style={at={(0.06,0.5)}}, |
|
269 enlargelimits=false, |
|
270 xtick={0,5,...,30}, |
|
271 xmax=33, |
|
272 ymax=45, |
|
273 ytick={0,10,...,40}, |
|
274 scaled ticks=false, |
|
275 axis lines=left, |
|
276 width=6cm, |
|
277 height=4.5cm, |
|
278 legend entries={Python, Java 8}, |
|
279 legend pos=north west] |
|
280 \addplot[blue,mark=*, mark options={fill=white}] table {re-python2.data}; |
|
281 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java.data}; |
|
282 \end{axis} |
|
283 \end{tikzpicture} |
|
284 \begin{tikzpicture} |
|
285 \begin{axis}[ |
|
286 xlabel={$n$}, |
|
287 x label style={at={(1.05,0.0)}}, |
|
288 ylabel={time in secs}, |
|
289 y label style={at={(0.06,0.5)}}, |
|
290 %enlargelimits=false, |
|
291 %xtick={0,5000,...,30000}, |
|
292 xmax=65000, |
|
293 ymax=45, |
|
294 ytick={0,10,...,40}, |
|
295 scaled ticks=false, |
|
296 axis lines=left, |
|
297 width=6cm, |
|
298 height=4.5cm, |
|
299 legend entries={Java 9}, |
|
300 legend pos=north west] |
|
301 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java9.data}; |
|
302 \end{axis} |
|
303 \end{tikzpicture} |
|
304 \end{column} |
|
305 \end{frame} |
|
306 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
307 |
|
308 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
309 \begin{frame}[c] |
|
310 \frametitle{Evil Regular Expressions} |
|
311 |
|
312 \begin{itemize} |
|
313 \item \alert{R}egular \alert{e}xpression \alert{D}enial \alert{o}f \alert{S}ervice (ReDoS)\medskip |
|
314 \item Evil regular expressions\medskip |
|
315 \begin{itemize} |
|
316 \item \bl{$(a^{?\{n\}}) \cdot a^{\{n\}}$} |
|
317 \item \bl{$(a^*)^*\cdot b$} |
|
318 \item \bl{$([a$\,-\,$z]^+)^*$} |
|
319 \item \bl{$(a + a \cdot a)^*$} |
|
320 \item \bl{$(a + a^?)^*$} |
|
321 \end{itemize} |
|
322 |
|
323 \item sometimes also called \alert{catastrophic backtracking} |
|
324 \item this is a problem for \alert{N}etwork \alert{I}ntrusion |
|
325 \alert{D}etection systems, StackExchange, Atom editor |
|
326 \item \url{https://vimeo.com/112065252} |
|
327 \end{itemize} |
|
328 |
|
329 \end{frame} |
|
330 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
331 |
|
332 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
333 \begin{frame}[c] |
|
334 \frametitle{The Goal of this Module} |
|
335 |
|
336 \begin{center} |
|
337 \begin{tikzpicture}[scale=1, |
|
338 node/.style={ |
|
339 rectangle,rounded corners=3mm, |
|
340 very thick,draw=black!50,minimum height=18mm, minimum width=20mm, |
|
341 top color=white,bottom color=black!20}] |
|
342 |
|
343 \node at (3.05, 1.8) {\Large\bf write a compiler}; |
|
344 |
|
345 \node (0) at (-2.3,0) {}; |
|
346 \node [above=5mm of 0] |
|
347 {\makebox[0mm]{\footnotesize |
|
348 \begin{tabular}{@{}l@{}}input\\[-1mm]program\end{tabular}}}; |
|
349 |
|
350 \node (A) at (0,0) [node] {}; |
|
351 \node [below right] at (A.north west) {lexer}; |
|
352 |
|
353 \node (B) at (3,0) [node] {}; |
|
354 \node [below right=1mm] at (B.north west) {\mbox{}\hspace{-1mm}parser}; |
|
355 |
|
356 \node (C) at (6,0) [node] {}; |
|
357 \node [below right] at (C.north west) {\mbox{}\hspace{-1mm}code gen}; |
|
358 |
|
359 \node (1) at (8.4,0) {}; |
|
360 \node [above=5mm of 1] |
|
361 {\makebox[0mm]{\footnotesize |
|
362 \begin{tabular}{@{}r@{}}binary\\[-1mm]code\end{tabular}}}; |
|
363 |
|
364 \draw [->,line width=4mm] (0) -- (A); |
|
365 \draw [->,line width=4mm] (A) -- (B); |
|
366 \draw [->,line width=4mm] (B) -- (C); |
|
367 \draw [->,line width=4mm] (C) -- (1); |
|
368 \end{tikzpicture} |
|
369 \end{center} |
|
370 |
|
371 \only<2,3,4>{ |
|
372 \begin{textblock}{1}(1,2.1) |
|
373 \begin{bubble}[9.8cm] |
|
374 \normalsize |
|
375 lexer input: a string\smallskip\\ |
|
376 \hspace{5mm}\code{"read(n);"}\medskip\\ |
|
377 lexer output: a sequence of tokens\smallskip\\ |
|
378 \hspace{5mm}\code{key(read) lpar id(n) rpar semi} |
|
379 \end{bubble} |
|
380 \end{textblock}} |
|
381 |
|
382 \only<3,4>{ |
|
383 \begin{textblock}{1}(6,7.8) |
|
384 \begin{tabular}{c} |
|
385 \includegraphics[scale=0.2]{rosetta.jpg}\\[-2mm] |
|
386 \footnotesize lexing $\Rightarrow$ recognising words (Stone of Rosetta) |
|
387 \end{tabular} |
|
388 \end{textblock}} |
|
389 |
|
390 \only<4>{ |
|
391 \begin{textblock}{1}(0.5,12)\small |
|
392 \begin{tabular}{l@{}c@{}l} |
|
393 \pcode{if} & $\;\Rightarrow\;$ & keyword\\ |
|
394 \pcode{iffoo} & $\;\Rightarrow\;$ & identifier\\ |
|
395 \end{tabular} |
|
396 \end{textblock}} |
|
397 |
|
398 \only<5>{ |
|
399 \begin{textblock}{1}(1,1.5) |
|
400 \begin{bubble}[8.5cm] |
|
401 \normalsize |
|
402 parser input: a sequence of tokens\smallskip\\ |
|
403 |
|
404 {\small\hspace{5mm}\code{key(read) lpar id(n) rpar semi}}\smallskip\\ |
|
405 |
|
406 parser output: an abstract syntax tree\smallskip\\ |
|
407 \footnotesize |
|
408 \hspace{2cm}\begin{tikzpicture} |
|
409 \node {\code{read}} |
|
410 child {node {\code{lpar}}} |
|
411 child {node {\code{n}}} |
|
412 child {node {\code{rpar}}}; |
|
413 \end{tikzpicture} |
|
414 \end{bubble} |
|
415 \end{textblock}} |
|
416 |
|
417 \only<6,7>{ |
|
418 \begin{textblock}{1}(1,1.5) |
|
419 \begin{bubble}[4cm] |
|
420 \normalsize |
|
421 code generator:\smallskip\\ |
|
422 \hspace{5mm}\code{istore 2}\\ |
|
423 \hspace{5mm}\code{iload 2}\\ |
|
424 \hspace{5mm}\code{ldc 10}\\ |
|
425 \hspace{5mm}\code{isub}\\ |
|
426 \hspace{5mm}\code{ifeq Label2}\\ |
|
427 \hspace{5mm}\code{iload 2}\\ |
|
428 \hspace{5mm}\code{...}\\ |
|
429 \end{bubble} |
|
430 \end{textblock}} |
|
431 |
|
432 \only<7>{ |
|
433 \begin{textblock}{6}(8.4,7) |
|
434 \begin{bubble}[5cm] |
|
435 \mbox{\begin{tikzpicture}[scale=0.58,rounded corners=0mm] |
|
436 \begin{axis}[axis x line=bottom, axis y line=left, ylabel=secs, |
|
437 xlabel=n, |
|
438 enlargelimits=0.05, |
|
439 ybar interval=0.7, legend style=small] |
|
440 \addplot file {interpreted2.data}; |
|
441 \addplot file {compiled2.data}; |
|
442 %\legend{interpreted, compiled} |
|
443 \end{axis} |
|
444 \end{tikzpicture}} |
|
445 \end{bubble} |
|
446 \end{textblock}} |
|
447 |
|
448 \end{frame} |
|
449 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
450 |
|
451 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
452 \begin{frame}[c] |
|
453 \frametitle{The Acad.~Subject is Mature} |
|
454 |
|
455 \begin{itemize} |
|
456 \item Turing Machines, 1936 |
|
457 \item Regular Expressions, 1956\\ |
|
458 \item The first compiler for COBOL, 1957\\ (Grace Hopper) |
|
459 \item But surprisingly research papers are still published nowadays\\ |
|
460 \item ``Parsing: The Solved Problem That Isn't'' |
|
461 \end{itemize} |
|
462 |
|
463 \begin{flushright} |
|
464 \includegraphics[scale=0.3]{hopper.jpg}\\ |
|
465 \footnotesize\textcolor{gray}{Grace Hopper} |
|
466 \end{flushright} |
|
467 |
|
468 |
|
469 \begin{flushright} |
|
470 \mbox{}\\[-6mm] |
|
471 {\footnotesize\textcolor{gray}{(she made it to David Letterman's Tonight Show,\\[-2mm] |
|
472 \url{http://www.youtube.com/watch?v=aZOxtURhfEU})}} |
|
473 \end{flushright} |
|
474 |
|
475 \end{frame} |
|
476 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
477 |
|
478 |
|
479 |
|
480 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
481 \begin{frame}[c] |
|
482 \frametitle{Lectures 1 - 5} |
|
483 |
|
484 transforming strings into structured data\\[10mm] |
|
485 |
|
486 {\LARGE\bf Lexing} {\hfill{}based on regular expressions}\medskip\\ |
|
487 \hspace{5mm}(recognising ``words'')\\[6mm] |
|
488 |
|
489 {\LARGE\bf Parsing}\medskip\\ |
|
490 \hspace{5mm}(recognising ``sentences'') |
|
491 |
|
492 \begin{textblock}{1}(10,9.1) |
|
493 \begin{tabular}{c} |
|
494 \includegraphics[scale=0.1]{rosetta.jpg}\\[-2mm] |
|
495 \footnotesize Stone of Rosetta |
|
496 \end{tabular} |
|
497 \end{textblock} |
|
498 |
|
499 \end{frame} |
|
500 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
501 |
|
502 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
503 \begin{frame}[t] |
|
504 \frametitle{Familiar Regular Expr.} |
|
505 \small |
|
506 \begin{center} |
|
507 \texttt{[a-z0-9\_$\backslash{}$.-]+ @ [a-z0-9$\backslash{}$.-]+ . [a-z$\backslash{}$.]\{2,6\}} |
|
508 \end{center}\smallskip |
|
509 |
|
510 \begin{center} |
|
511 \begin{tabular}{@{}lp{8.5cm}@{}} |
|
512 \pcode{re*} & matches 0 or more times\\ |
|
513 \pcode{re+} & matches 1 or more times\\ |
|
514 \pcode{re?} & matches 0 or 1 times\\ |
|
515 \pcode{re\{n\}} & matches exactly \pcode{n} number of times\\ |
|
516 \pcode{re\{n,m\}} & matches at least \pcode{n} and at most {\tt m} times\\ |
|
517 \pcode{[...]} & matches any single character inside the brackets\\ |
|
518 \pcode{[^...]} & matches any single character not inside the |
|
519 brackets\\ |
|
520 \pcode{a-z A-Z} & character ranges\\ |
|
521 \pcode{\\d} & matches digits; equivalent to \pcode{[0-9]}\\ |
|
522 \pcode{.} & matches every character except newline\\ |
|
523 \pcode{(re)} & groups regular expressions and remembers |
|
524 the matched text |
|
525 \end{tabular} |
|
526 \end{center} |
|
527 |
|
528 |
|
529 \end{frame} |
|
530 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
531 |
|
532 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
533 \begin{frame}[c] |
|
534 \frametitle{Today} |
|
535 |
|
536 \begin{itemize} |
|
537 \item While the ultimate goal is to implement a small compiler for the JVM |
|
538 \ldots\bigskip |
|
539 \end{itemize} |
|
540 |
|
541 Let's start with: |
|
542 |
|
543 \begin{itemize} |
|
544 \item a web-crawler |
|
545 \item an email harvester |
|
546 %\item \textcolor{gray}{(a web-scraper)} |
|
547 \end{itemize} |
|
548 |
|
549 \end{frame} |
|
550 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
551 |
|
552 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
553 \begin{frame}[t] |
|
554 \frametitle{A Web-Crawler} |
|
555 |
|
556 \mbox{}\\[10mm] |
|
557 |
|
558 \begin{enumerate} |
|
559 \item given an URL, read the corresponding webpage |
|
560 \item extract all links from it |
|
561 \item call the web-crawler again for all these links |
|
562 \end{enumerate} |
|
563 |
|
564 \end{frame} |
|
565 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
566 |
|
567 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
568 \begin{frame}[t] |
|
569 \frametitle{A Web-Crawler} |
|
570 |
|
571 \mbox{}\\[10mm] |
|
572 |
|
573 |
|
574 \begin{enumerate} |
|
575 \item given an URL, read the corresponding webpage |
|
576 \item if not possible print, out a problem |
|
577 \item if possible, extract all links from it |
|
578 \item call the web-crawler again for all these links |
|
579 \end{enumerate}\bigskip\pause |
|
580 |
|
581 \small (we need a bound for the number of recursive calls) |
|
582 |
|
583 \small (the purpose is to check all links on my own webpage) |
|
584 \end{frame} |
|
585 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
586 |
|
587 |
|
588 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
589 \begin{frame}[c] |
|
590 |
|
591 \begin{textblock}{1}(2,5) |
|
592 \begin{tabular}{c} |
|
593 \includegraphics[scale=0.15]{servers.png}\\[-2mm] |
|
594 \small Server |
|
595 \end{tabular} |
|
596 \end{textblock} |
|
597 |
|
598 \begin{textblock}{1}(5.6,4) |
|
599 \begin{tikzpicture}[scale=1.1] |
|
600 \draw[white] (0,1) node (X) {}; |
|
601 \draw[white] (2,1) node (Y) {}; |
|
602 \draw[white] (0,0) node (X1) {}; |
|
603 \draw[white] (2,0) node (Y1) {}; |
|
604 \draw[white] (0,-1) node (X2) {}; |
|
605 \draw[white] (2,-1) node (Y2) {}; |
|
606 \draw[red, <-, line width = 2mm] (X) -- (Y); |
|
607 \node [inner sep=5pt,label=above:\textcolor{black}{GET request}] at ($ (X)!.5!(Y) $) {}; |
|
608 \draw[red, ->, line width = 2mm] (X1) -- (Y1); |
|
609 \node [inner sep=5pt,label=above:\textcolor{black}{webpage}] at ($ (X1)!.5!(Y1) $) {}; |
|
610 \draw[red, <-, line width = 2mm] (X2) -- (Y2); |
|
611 \node [inner sep=7pt,label=above:\textcolor{black}{POST data}] at ($ (X2)!.5!(Y2) $) {}; |
|
612 \end{tikzpicture} |
|
613 \end{textblock} |
|
614 |
|
615 |
|
616 \begin{textblock}{1}(9,5.5) |
|
617 \begin{tabular}{c} |
|
618 \includegraphics[scale=0.15]{laptop.png}\\[-2mm] |
|
619 \small Browser |
|
620 \end{tabular} |
|
621 \end{textblock} |
|
622 \end{frame} |
|
623 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
624 |
|
625 |
|
626 |
|
627 |
|
628 |
|
629 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
630 \begin{frame}[t] |
|
631 \frametitle{A Regular Expression} |
|
632 |
|
633 \begin{itemize} |
|
634 \item \ldots{} is a pattern or template for specifying strings |
|
635 \end{itemize}\bigskip |
|
636 |
|
637 \begin{center} |
|
638 \only<1>{\scode{"https?://[^"]*"}}% |
|
639 \only<2>{\scode{""""https?://[^"]*"""".r}} |
|
640 \end{center}\bigskip\bigskip |
|
641 |
|
642 matches for example\smallskip\\ |
|
643 \hspace{2mm}\code{"http://www.foobar.com"}\\ |
|
644 \hspace{2mm}\code{"https://www.tls.org"}\smallskip\\ |
|
645 |
|
646 but not\smallskip\\ |
|
647 \hspace{2mm}\code{"http://www."foo"bar.com"}\\ |
|
648 |
|
649 \end{frame} |
|
650 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
651 |
|
652 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
653 \begin{frame}[c] |
|
654 \frametitle{Finding Operations in Scala} |
|
655 |
|
656 {\bf\code{rexp.findAllIn(string)}}\medskip |
|
657 |
|
658 returns a list of all (sub)strings that match the |
|
659 regular expression |
|
660 \bigskip\bigskip |
|
661 |
|
662 |
|
663 {\bf\code{rexp.findFirstIn(string)}}\medskip |
|
664 |
|
665 returns either |
|
666 |
|
667 \begin{itemize} |
|
668 \item \code{None} if no (sub)string matches or |
|
669 \item \code{Some(s)} with the first (sub)string |
|
670 \end{itemize} |
|
671 |
|
672 \end{frame} |
|
673 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
674 |
|
675 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
676 \begin{frame}[t] |
196 \begin{frame}[t] |
677 \frametitle{Regular Expressions} |
197 \frametitle{Regular Expressions} |
678 |
198 |
679 Their inductive definition: |
199 Their inductive definition:\\ |
680 |
200 \hspace{10pt} |
681 |
201 \vspace{100pt} |
682 \begin{textblock}{6}(2,7.5) |
202 |
|
203 \begin{textblock}{6}(1.5,3.5) |
683 \begin{tabular}{@ {}rrl@ {\hspace{13mm}}l} |
204 \begin{tabular}{@ {}rrl@ {\hspace{13mm}}l} |
684 \bl{$r$} & \bl{$::=$} & \bl{$\ZERO$} & nothing\\ |
205 \bl{$r$} & \bl{$::=$} & \bl{$\ZERO$} & nothing\\ |
685 & \bl{$\mid$} & \bl{$\ONE$} & empty string / \pcode{""} / $[]$\\ |
206 & \bl{$\mid$} & \bl{$\ONE$} & empty string / \pcode{""} / $[]$\\ |
686 & \bl{$\mid$} & \bl{$c$} & character\\ |
207 & \bl{$\mid$} & \bl{$c$} & character\\ |
687 & \bl{$\mid$} & \bl{$r_1 + r_2$} & alternative / choice\\ |
208 & \bl{$\mid$} & \bl{$r_1 + r_2$} & alternative / choice\\ |
688 & \bl{$\mid$} & \bl{$r_1 \cdot r_2$} & sequence\\ |
209 & \bl{$\mid$} & \bl{$r_1 \cdot r_2$} & sequence\\ |
689 & \bl{$\mid$} & \bl{$r^*$} & star (zero or more)\\ |
210 & \bl{$\mid$} & \bl{$r^*$} & star (zero or more)\\ |
690 \end{tabular} |
211 \end{tabular} |
691 \end{textblock} |
212 \end{textblock} |
692 |
213 |
|
214 \begin{itemize} |
|
215 \item { Formalized by Stephen Coles Kleene in 1950s} |
|
216 \item {\bf What could be possibly interesting?} |
|
217 \end{itemize} |
|
218 \end{frame} |
|
219 |
|
220 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
221 \iffalse |
|
222 \begin{frame}[t] |
|
223 \frametitle{Regular Expressions} |
|
224 |
|
225 John Regehr {\small(Univ.~Utah, LLVM compiler hacker)}\smallskip\\ |
|
226 |
|
227 \begin{bubble}[10.5cm] |
|
228 \bf ``\ldots{}It’s effectively a perpetual |
|
229 employment act for solid compiler hackers.'' |
|
230 \end{bubble} |
|
231 |
|
232 \onslide<1->{ |
|
233 \only<2>{ |
|
234 \begin{itemize} |
|
235 \item {\bf Hardware is getting weirder |
|
236 rather than getting clocked faster} |
|
237 |
|
238 \begin{itemize} |
|
239 \item Almost all processors are |
|
240 multicores nowadays and it looks like there is increasing asymmetry in |
|
241 resources across cores. Processors come with vector units, crypto |
|
242 accelerators etc. We have DSPs, GPUs, |
|
243 ARM big.little, and Xeon Phi. This is only scratching the |
|
244 surface. |
|
245 \end{itemize} |
|
246 \end{itemize}} |
|
247 \only<3>{ |
|
248 \begin{itemize} |
|
249 \item {\bf We’re getting tired of low-level languages and |
|
250 their associated security disasters} |
693 |
251 |
694 |
252 \begin{itemize} |
695 \end{frame} |
253 \item |
696 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
254 We want to write new code, to |
697 |
255 whatever extent possible, in safer, higher-level |
698 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
256 languages. Compilers are caught right in the middle of these |
699 %\begin{frame}[t] |
257 opposing trends: one of their main jobs is to help bridge the large |
700 %\frametitle{Regular Expressions} |
258 and growing gap between increasingly high-level languages and |
701 % |
259 increasingly wacky platforms. |
702 %\small |
260 \end{itemize} |
703 %In Scala:\bigskip |
261 \end{itemize}}} |
704 % |
262 |
705 %\footnotesize |
263 \end{frame} |
706 %\lstinputlisting{../progs/app51.scala} |
264 \fi |
707 % |
265 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
708 % |
266 |
709 %\end{frame} |
267 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
710 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
268 \iffalse |
711 |
|
712 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
713 \begin{frame}[t] |
|
714 \frametitle{Strings} |
|
715 |
|
716 \ldots are lists of characters. For example \code{"hello"} |
|
717 |
|
718 \begin{center} |
|
719 \bl{$[h, e, l, l, o]$} or just \bl{$hello$} |
|
720 \end{center} |
|
721 |
|
722 the empty string: \bl{$[]$} or \bl{\pcode{""}}\bigskip\\ |
|
723 |
|
724 the concatenation of two strings: |
|
725 |
|
726 \begin{center} |
|
727 \bl{$s_1 \,@\, s_2$} |
|
728 \end{center} |
|
729 |
|
730 \bl{\textit{foo $@$ bar = foobar}, \textit{baz $@\, []$ = baz}} |
|
731 |
|
732 \end{frame} |
|
733 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
734 |
|
735 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
736 \begin{frame}[c] |
269 \begin{frame}[c] |
737 \frametitle{Languages, Strings} |
270 \frametitle{Why Bother?} |
738 |
|
739 \begin{itemize} |
|
740 \item \alert{\bf Strings} are lists of characters, for example |
|
741 \begin{center} |
|
742 \bl{$[]$},\;\bl{$abc$} \hspace{2cm}(Pattern match: \bl{$c\!::\!s$}) |
|
743 \end{center}\bigskip |
|
744 |
|
745 |
|
746 \item A \alert{\bf language} is a set of strings, for example\medskip |
|
747 \begin{center} |
|
748 \bl{$\{[], hello, \textit{foobar}, a, abc\}$} |
|
749 \end{center}\bigskip |
|
750 |
|
751 \item \alert{\bf Concatenation} of strings and languages |
|
752 |
|
753 \begin{center} |
|
754 \begin{tabular}{rcl} |
|
755 \bl{$\textit{foo}\;@\;bar$} & \bl{$=$} & \bl{$\textit{foobar}$}\medskip\\ |
|
756 \bl{$A\;@\;B$} & \bl{$\dn$} & \bl{$\{ s_1\,@\,s_2 \;\mid\; s_1 \in A \wedge s_2 \in B\}$} |
|
757 \end{tabular} |
|
758 \end{center} |
|
759 |
|
760 %\item The \alert{\bf meaning} of a regular expression is a set of |
|
761 % strings, or language. |
|
762 \end{itemize} |
|
763 |
|
764 \end{frame} |
|
765 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
766 |
|
767 |
|
768 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
769 \mode<presentation>{ |
|
770 \begin{frame}[c] |
|
771 \frametitle{\begin{tabular}{c}The Meaning of a\\[-2mm] |
|
772 Regular Expression\end{tabular}} |
|
773 |
|
774 \begin{textblock}{15}(1,4) |
|
775 \begin{tabular}{rcl} |
|
776 \bl{$L(\ZERO)$} & \bl{$\dn$} & \bl{$\{\}$}\\ |
|
777 \bl{$L(\ONE)$} & \bl{$\dn$} & \bl{$\{[]\}$}\\ |
|
778 \bl{$L(c)$} & \bl{$\dn$} & \bl{$\{[c]\}$}\\ |
|
779 \bl{$L(r_1 + r_2)$} & \bl{$\dn$} & \bl{$L(r_1) \cup L(r_2)$}\\ |
|
780 \bl{$L(r_1 \cdot r_2)$} & \bl{$\dn$} & \bl{$\{ s_1 \,@\, s_2 \;|\; s_1 \in L(r_1) \wedge s_2 \in L(r_2) \}$}\\ |
|
781 \bl{$L(r^*)$} & \bl{$\dn$} & \onslide<4->{\bl{$\bigcup_{0 \le n} L(r)^n$}}\\ |
|
782 \end{tabular}\bigskip |
|
783 |
|
784 \onslide<2->{ |
|
785 \hspace{5mm}\bl{$L(r)^0 \;\dn\; \{[]\}$}\\ |
|
786 \bl{$L(r)^{n+1} \;\dn\; L(r) \,@\, L(r)^n$}\hspace{9mm}\onslide<3->{\small\textcolor{gray}{(append on sets)}\\ |
|
787 \small\hspace{5cm}\textcolor{gray}{$\{ s_1 @ s_2 \;|\; s_1\in L(r) \wedge s_2 \in L(r)^n \}$}} |
|
788 } |
|
789 \end{textblock} |
|
790 |
|
791 \end{frame}} |
|
792 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
793 |
|
794 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
795 \begin{frame}[c] |
|
796 \frametitle{The Meaning of Matching} |
|
797 |
|
798 \begin{bubble}[10cm] |
|
799 \large\bf |
|
800 A regular expression \bl{$r$} matches a string~\bl{$s$} |
|
801 provided |
|
802 |
|
803 \begin{center} |
|
804 \bl{$s \in L(r)$}\\ |
|
805 \end{center} |
|
806 \end{bubble}\bigskip\bigskip |
|
807 |
|
808 \ldots and the point of the next lecture is |
|
809 to decide this problem as fast as possible (unlike Python, |
|
810 Ruby, Java) |
|
811 |
|
812 \end{frame} |
|
813 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
814 |
|
815 |
|
816 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
817 \begin{frame}[c] |
|
818 \frametitle{Written Exam} |
|
819 |
|
820 \begin{itemize} |
|
821 \item Accounts for 80\%.\bigskip |
|
822 |
|
823 \item The question ``\textit{Is this relevant for |
|
824 the exam?}'' is very demotivating for the lecturer!\bigskip\\ |
|
825 |
|
826 \item Deal: Whatever is in the homework (and is not marked |
|
827 ``\textit{optional}'') is relevant for the exam.\bigskip |
|
828 |
|
829 \item Each lecture has also a handout. There are also handouts about |
|
830 notation and Scala. |
|
831 \end{itemize} |
|
832 |
|
833 \end{frame} |
|
834 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
835 |
|
836 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
837 \begin{frame}[t] |
|
838 \frametitle{Coursework} |
|
839 |
|
840 \begin{itemize} |
|
841 \item Accounts for 20\%. Two strands. Choose \alert{\bf one}!\bigskip |
|
842 \end{itemize} |
|
843 |
271 |
844 \begin{columns}[t] |
272 \begin{columns}[t] |
845 \begin{column}{.5\textwidth} |
273 \begin{column}{.5\textwidth} |
846 \underline{\bf Strand 1}\medskip |
274 Ruby, Python, Java 8\medskip\\ |
847 \begin{itemize} |
275 \begin{tikzpicture}\footnotesize |
848 \item four programming tasks: |
276 \begin{axis}[ |
849 \begin{itemize} |
277 xlabel={$n$}, |
850 \item matcher (4\%, 12.10.) |
278 x label style={at={(1.05,0.0)}}, |
851 \item lexer (5\%, 02.11.) |
279 ylabel={time in secs}, |
852 \item parser (5\%, 23.11.) |
280 enlargelimits=false, |
853 \item compiler (6\%, 14.12.) |
281 xtick={0,5,...,30}, |
854 \end{itemize} |
282 xmax=33, |
855 \item in any lang.~you like,\\ but I want to see the code |
283 ymax=35, |
856 \end{itemize} |
284 ytick={0,5,...,30}, |
|
285 scaled ticks=false, |
|
286 axis lines=left, |
|
287 width=5.5cm, |
|
288 height=4cm, |
|
289 legend entries={Python,Ruby}, |
|
290 legend pos=north west, |
|
291 legend cell align=left] |
|
292 \addplot[blue,mark=*, mark options={fill=white}] table {re-python.data}; |
|
293 \addplot[brown,mark=triangle*, mark options={fill=white}] table {re-ruby.data}; |
|
294 \end{axis} |
|
295 \end{tikzpicture} |
|
296 \begin{tikzpicture}\footnotesize |
|
297 \begin{axis}[ |
|
298 xlabel={$n$}, |
|
299 x label style={at={(1.05,0.0)}}, |
|
300 ylabel={time in secs}, |
|
301 enlargelimits=false, |
|
302 xtick={0,5,...,30}, |
|
303 xmax=33, |
|
304 ymax=35, |
|
305 ytick={0,5,...,30}, |
|
306 scaled ticks=false, |
|
307 axis lines=left, |
|
308 width=5.5cm, |
|
309 height=4cm, |
|
310 legend entries={Python, Java 8}, |
|
311 legend pos=north west, |
|
312 legend cell align=left] |
|
313 \addplot[blue,mark=*, mark options={fill=white}] table {re-python2.data}; |
|
314 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java.data}; |
|
315 \end{axis} |
|
316 \end{tikzpicture} |
|
317 |
857 \end{column} |
318 \end{column} |
858 |
|
859 \hspace{-45pt}\vrule{}\hspace{10pt} |
|
860 \begin{column}{.5\textwidth} |
319 \begin{column}{.5\textwidth} |
861 \underline{\bf Strand 2}\smallskip\begin{itemize} |
320 Us (after next lecture)\medskip\\ |
862 \item one task: prove the correctness of a regular expression matcher in |
321 \begin{tikzpicture}\footnotesize |
863 the \underline{Isabelle} theorem prover |
322 \begin{axis}[ |
864 \item 20\%, submission on~14.12.\hspace{-5mm}\mbox{} |
323 xlabel={$n$}, |
865 \end{itemize} |
324 x label style={at={(1.07,0.0)}}, |
|
325 ylabel={time in secs}, |
|
326 enlargelimits=false, |
|
327 xtick={0,5000,...,10000}, |
|
328 xmax=11000, |
|
329 ymax=35, |
|
330 ytick={0,5,...,30}, |
|
331 scaled ticks=false, |
|
332 axis lines=left, |
|
333 width=5.5cm, |
|
334 height=4cm] |
|
335 \addplot[green,mark=square*,mark options={fill=white}] table {re2.data}; |
|
336 \addplot[black,mark=square*,mark options={fill=white}] table {re3.data}; |
|
337 \end{axis} |
|
338 \end{tikzpicture} |
|
339 \begin{tikzpicture}\footnotesize |
|
340 \begin{axis}[ |
|
341 xlabel={$n$}, |
|
342 x label style={at={(1.07,0.0)}}, |
|
343 ylabel={time in secs}, |
|
344 enlargelimits=false, |
|
345 ymax=35, |
|
346 ytick={0,5,...,30}, |
|
347 scaled ticks=false, |
|
348 axis lines=left, |
|
349 width=5.5cm, |
|
350 height=4cm] |
|
351 \addplot[black,mark=square*,mark options={fill=white}] table {re3a.data}; |
|
352 \end{axis} |
|
353 \end{tikzpicture} |
866 \end{column} |
354 \end{column} |
867 \end{columns}\medskip |
355 \end{columns}\bigskip |
868 |
356 |
869 \small |
357 \small\centering |
870 \begin{itemize} |
358 matching \texttt{[a?]\{n\}[a]\{n\}} and \texttt{(a*)*b} |
871 \item Solving more than one strand will {\bf not} give you more |
359 against $\underbrace{\texttt{a}...\texttt{a}}_n$ |
872 marks. |
360 \end{frame} |
873 |
361 \fi |
874 \end{itemize} |
362 |
875 |
363 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
876 \end{frame} |
364 |
877 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
365 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
878 |
366 |
879 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
880 \begin{frame}[c] |
367 \begin{frame}[c] |
881 \frametitle{Lecture Capture} |
368 \frametitle{Catastrophic Backtracking} |
882 |
369 |
883 \begin{itemize} |
370 \begin{columns} |
884 \item Hope it works\ldots\pause actually no, it does not!\medskip\pause |
371 \begin{column}{.5\textwidth} |
885 \item It is important to use lecture capture wisely\\ (it is only the ``baseline''): |
372 |
886 \begin{itemize} |
373 \begin{tikzpicture} |
887 \item Lecture recordings are a study and revision aid. |
374 \begin{axis}[ |
888 \item Statistically, there is a clear and direct link between attendance and |
375 xlabel={$n$}, |
889 attainment: Students who do not attend lectures, do less well in exams. |
376 x label style={at={(1.07,0.0)}}, |
890 \end{itemize} |
377 ylabel={time in secs}, |
891 |
378 %y label style={at={(0.06,0.5)}}, |
892 \item Attending a lecture is more than watching it online -- if you do not |
379 enlargelimits=false, |
893 attend, you miss out! |
380 xtick={0,5,...,30}, |
894 |
381 xmax=33, |
895 \end{itemize} |
382 ymax=45, |
896 |
383 ytick={0,10,...,40}, |
897 \end{frame} |
384 scaled ticks=false, |
898 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
385 axis lines=left, |
|
386 width=5cm, |
|
387 height=4cm, |
|
388 legend entries={Python, Java 8}, |
|
389 legend pos=north west] |
|
390 \addplot[blue,mark=*, mark options={fill=white}] table {re-python2.data}; |
|
391 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java.data}; |
|
392 \end{axis} |
|
393 \end{tikzpicture} |
|
394 |
|
395 \begin{tikzpicture} |
|
396 \begin{axis}[ |
|
397 xlabel={$n$}, |
|
398 x label style={at={(1.07,0.0)}}, |
|
399 ylabel={time in secs}, |
|
400 %y label style={at={(0.06,0.5)}}, |
|
401 %enlargelimits=false, |
|
402 %xtick={0,5000,...,30000}, |
|
403 xmax=65000, |
|
404 ymax=45, |
|
405 ytick={0,10,...,40}, |
|
406 scaled ticks=false, |
|
407 axis lines=left, |
|
408 width=5cm, |
|
409 height=4cm, |
|
410 legend entries={Java 9}, |
|
411 legend pos=north west] |
|
412 \addplot[cyan,mark=*, mark options={fill=white}] table {re-java9.data}; |
|
413 \end{axis} |
|
414 \end{tikzpicture} |
|
415 \end{column} |
|
416 \end{columns}\bigskip |
|
417 \center |
|
418 matching \texttt{(a*)*b} |
|
419 against $\underbrace{\texttt{a}...\texttt{a}}_n$ |
|
420 \end{frame} |
|
421 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
422 |
|
423 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
424 \begin{frame}[fragile] |
|
425 \frametitle{Impact} |
|
426 |
|
427 \begin{itemize} |
|
428 \item Network Traffic Analysis |
|
429 \begin{itemize} |
|
430 \item Snort: > 5 million downloads |
|
431 \item Bro: > 10000 downloads |
|
432 \item thousands of regexes |
|
433 \item \alert{R}egular \alert{e}xpression \alert{D}enial \alert{o}f \alert{S}ervice (ReDoS)\medskip |
|
434 |
|
435 \end{itemize} |
|
436 |
|
437 \begin{verbatim} |
|
438 Jan 2 00:53:19 talisker sshd: Received disconnect from 110.53.183.227: [preauth] |
|
439 Jan 2 00:58:31 talisker sshd: Received disconnect from 110.53.183.252: [preauth] |
|
440 Jan 2 01:01:28 talisker sshd: Received disconnect from 221.194.47.236: [preauth] |
|
441 Jan 2 01:03:59 talisker sshd: Received disconnect from 110.53.183.228: [preauth] |
|
442 Jan 2 01:06:53 talisker sshd: Received disconnect from 221.194.47.245: [preauth] |
|
443 ... |
|
444 \end{verbatim} |
|
445 |
|
446 |
|
447 \item Compilers: lexical analysis |
|
448 \end{itemize} |
|
449 |
|
450 \end{frame} |
|
451 |
|
452 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
453 |
|
454 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
455 \begin{frame}[c] |
|
456 \frametitle{Evil Regular Expressions} |
|
457 |
|
458 \begin{itemize} |
|
459 %\item \alert{R}egular \alert{e}xpression \alert{D}enial \alert{o}f \alert{S}ervice (ReDoS)\medskip |
|
460 \item Evil regular expressions\medskip |
|
461 \begin{itemize} |
|
462 \item \bl{$(a^{?\{n\}}) \cdot a^{\{n\}}$} |
|
463 \item \bl{$(a^*)^*\cdot b$} |
|
464 \item \bl{$([a$\,-\,$z]^+)^*$} |
|
465 \item \bl{$(a + a \cdot a)^*$} |
|
466 \item \bl{$(a + a^?)^*$} |
|
467 \item \bl{$\^(.*?,)\{11\}P$} |
|
468 \item \bl{$<html>.*?<head>.*?<title>.*?</title>.*?</head>.*?<body[^>]*>.*?</body>.*?</html>$} |
|
469 \end{itemize} |
|
470 |
|
471 \item Pearl Compatible Regular Expression(PCRE) \bl{$(r) \backslash 1$} |
|
472 \end{itemize} |
|
473 |
|
474 \end{frame} |
|
475 |
|
476 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
477 |
|
478 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
479 |
|
480 \begin{frame}[c] |
|
481 \frametitle{A Concrete Example} |
|
482 \begin{adjustwidth*}{}{-1.2cm} |
|
483 \resizebox{13cm}{4cm}{ |
|
484 |
|
485 \begin{tikzpicture}[shorten >=1pt,node distance=2cm,on grid,auto] |
|
486 \node[state,initial] (q_0) {$0$}; |
|
487 \node[state] (q_1) [right = of q_0] {$1$}; |
|
488 \node[state] (q_2) [right=of q_1] {$2$}; |
|
489 \node (q_dots1) [right = of q_2] {$\cdots$}; |
|
490 \node [state] (q_n) [right = of q_dots1] {$n$}; |
|
491 \node [state] (q_n1) [right = of q_n] {$n+1$}; |
|
492 \node[state] (q_n2) [right=of q_n1] {$n+2$}; |
|
493 \node[state,accepting] (q_n3) [right=of q_n2] {$n+3$}; |
|
494 |
|
495 \path[->] |
|
496 (q_0) edge [loop above] node {$*$} (q_0) |
|
497 (q_0) edge node {$a$} (q_1) |
|
498 (q_1) edge node {$*$} (q_2) |
|
499 (q_2) edge node {$*$} (q_dots1) |
|
500 (q_dots1) edge node{$*$} (q_n) |
|
501 (q_n) edge node{$*$} (q_n1) |
|
502 (q_n1) edge node{$*$} (q_n2) |
|
503 (q_n2) edge node{$*$} (q_n3); |
|
504 |
|
505 \draw [decorate,decoration={brace,amplitude=10pt,mirror,raise=10pt},yshift=0pt] |
|
506 (q_2.south west) -- (q_n1.south east) node [black,midway,yshift = -2cm] {\footnotesize n states}; |
|
507 %\draw [decorate,decoration={brace,amplitude=10pt,mirror,raise=4pt},yshift=0pt] |
|
508 %(3.5,0.65) -- (3.5,6.5) node [black,midway,xshift=0.8cm] {\footnotesize |
|
509 %$P_2$}; |
|
510 \end{tikzpicture} |
|
511 } |
|
512 \end{adjustwidth*} |
|
513 |
|
514 \begin{itemize} |
|
515 \item |
|
516 NFA accepting Regex $.*a.{n}bc$ |
|
517 \item |
|
518 Matching the string aaaaaaaaaaaaaaa.....abc |
|
519 \end{itemize} |
|
520 \end{frame} |
|
521 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
522 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
523 \begin{frame}[c] |
|
524 \frametitle{Brzozowski Derivative} |
|
525 |
|
526 \begin{itemize} |
|
527 \item Brzozowski invented in his 1964 Phd thesis |
|
528 \item Intuition: Chopping off the first character |
|
529 |
|
530 \begin{center} |
|
531 \bl{$A \backslash c \dn \{ s \;|\; c\!::\!s \in A\}$ } |
|
532 \end{center} |
|
533 \item |
|
534 For \bl{$A = \{\textit{foo}, \textit{bar}, \textit{frak}\}$} then |
|
535 |
|
536 \begin{center} |
|
537 \bl{\begin{tabular}{l@{\hspace{2mm}}c@{\hspace{2mm}}l} |
|
538 $A \backslash f $ & $=$ & $\{\textit{oo}, \textit{rak}\}$\\ |
|
539 $A \backslash b $ & $=$ & $\{\textit{ar}\}$\\ |
|
540 $A \backslash a $ & $=$ & $\{\}$\pause |
|
541 \end{tabular}} |
|
542 \end{center} |
|
543 \item |
|
544 \begin{tabular}{rp{4cm}} |
|
545 \bl {$r \,matches \,s=c_1c_2...c_n$} |
|
546 \bl{$\Leftrightarrow$} \bl{$c_1c_2...c_n \in L(r)$} \bl{$\Leftrightarrow$}\\ |
|
547 \bl{$c_2...c_n \in L(r) \backslash c_1$}\\ |
|
548 \bl{$\Leftrightarrow$} \bl{$\cdots$}\\ |
|
549 \bl{$\Leftrightarrow$} \bl{$[] \in ((L(r)\backslash c_1) \backslash c_2)\cdots\backslash c_n$} |
|
550 \end{tabular} |
|
551 |
|
552 \end{itemize} |
|
553 |
|
554 \end{frame} |
|
555 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
556 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
557 |
|
558 \begin{frame}[c] |
|
559 \frametitle{Previous Example} |
|
560 |
|
561 \begin{center} |
|
562 \begin{tikzpicture} |
|
563 \begin{axis}[ |
|
564 xlabel={$n$}, |
|
565 x label style={at={(1.09,-0.15)}}, |
|
566 ylabel={time in secs}, |
|
567 scaled x ticks=false, |
|
568 enlargelimits=false, |
|
569 xtick distance=10000, |
|
570 xmax=44000, |
|
571 ytick={0,10,...,30}, |
|
572 ymax=35, |
|
573 axis lines=left, |
|
574 width=7cm, |
|
575 height=3.5cm, |
|
576 legend entries={Java \liningnums{9}+}, |
|
577 legend pos=north west, |
|
578 legend cell align=left] |
|
579 \addplot[blue,mark=square*,mark options={fill=white}] table {re-java9.data}; |
|
580 \end{axis} |
|
581 \end{tikzpicture} |
|
582 \end{center} |
|
583 |
|
584 \begin{center} |
|
585 \begin{tikzpicture} |
|
586 \begin{axis}[ |
|
587 xlabel={$n$}, |
|
588 x label style={at={(1.09,0.0)}}, |
|
589 ylabel={time in secs}, |
|
590 scaled x ticks=false, |
|
591 xtick distance=2000000, |
|
592 enlargelimits=false, |
|
593 xmax=6400000, |
|
594 ytick={0,10,...,30}, |
|
595 ymax=35, |
|
596 axis lines=left, |
|
597 width=7cm, |
|
598 height=3.5cm, |
|
599 legend entries={Simple Scala}, |
|
600 legend pos=north west, |
|
601 legend cell align=left] |
|
602 %\addplot[green,mark=square*,mark options={fill=white}] table {re2a.data}; |
|
603 \addplot[black,mark=square*,mark options={fill=white}] table {re3a.data}; |
|
604 \end{axis} |
|
605 \end{tikzpicture} |
|
606 \end{center} |
|
607 |
|
608 Regex: \bl{$(a^*)^* \cdot b$} \space\space\space\space\space\space |
|
609 Strings of the form \bl{$\underbrace{\,a\ldots a\,}_{n}$} |
|
610 |
|
611 \end{frame} |
|
612 |
|
613 |
|
614 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
615 \begin{frame}[t] |
|
616 \frametitle{Optimization} |
|
617 |
|
618 \begin{center} |
|
619 \begin{tikzpicture} |
|
620 \begin{semilogyaxis}[ |
|
621 xlabel={$n$}, |
|
622 x label style={at={(1.05,0.0)}}, |
|
623 ylabel={regex size}, |
|
624 enlargelimits=false, |
|
625 xtick={1,4,...,30}, |
|
626 xmax=33, |
|
627 %ytick={0, 10,...,100}, |
|
628 scaled ticks=false, |
|
629 axis lines=left, |
|
630 width=9cm, |
|
631 height=4.5cm, |
|
632 legend entries={Simple Scala}, |
|
633 legend pos= south east, |
|
634 legend cell align=left |
|
635 ] |
|
636 \addplot[black,mark=square*, mark options={fill=white}] table {re-usize.data}; |
|
637 %\addplot[brown,mark=pentagon*, mark options={fill=white}] table {re-ssize.data}; |
|
638 %\addplot[red,mark=triangle*,mark options={fill=white}] table {re1.data}; |
|
639 %\addplot[green,mark=square*,mark options={fill=white}] table {re2.data}; |
|
640 \end{semilogyaxis} |
|
641 \end{tikzpicture} |
|
642 \end{center} |
|
643 Regex: \bl{$(a+aa)^* \cdot b$} \space\space\space\space\space\space |
|
644 Strings of the form \bl{$\underbrace{\,a\ldots a\,}_{n}$}\\ |
|
645 \begin{itemize} |
|
646 \item |
|
647 Solution: Simplification%(Sulzmann \& Lu) |
|
648 \begin{itemize} |
|
649 \item $r+(r+s) = r+r+s = r+s$ |
|
650 \item $1 \cdot r= r$ |
|
651 \end{itemize} |
|
652 \end{itemize} |
|
653 \end{frame} |
|
654 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
655 \begin{frame}[t] |
|
656 \frametitle{Optimization} |
|
657 |
|
658 Regex: \bl{$(a+aa)^* \cdot b$} \space\space\space\space\space\space |
|
659 Strings of the form \bl{$\underbrace{\,a\ldots a\,}_{n}$} |
|
660 |
|
661 \begin{center} |
|
662 \begin{tikzpicture} |
|
663 \begin{axis}[ |
|
664 xlabel={$n$}, |
|
665 x label style={at={(1.05,0.0)}}, |
|
666 ylabel={regex size}, |
|
667 enlargelimits=false, |
|
668 xtick={1,4,...,30}, |
|
669 xmax=33, |
|
670 %ytick={0, 10,...,100}, |
|
671 scaled ticks=false, |
|
672 axis lines=left, |
|
673 width=9cm, |
|
674 height=4.5cm, |
|
675 legend entries={Simplification Applied}, |
|
676 legend pos= south east, |
|
677 legend cell align=left |
|
678 ] |
|
679 %\addplot[black,mark=square*, mark options={fill=white}] table {re-usize.data}; |
|
680 \addplot[black,mark=triangle*, mark options={fill=white}] table {re-ssize.data}; |
|
681 %\addplot[red,mark=triangle*,mark options={fill=white}] table {re1.data}; |
|
682 %\addplot[green,mark=square*,mark options={fill=white}] table {re2.data}; |
|
683 \end{axis} |
|
684 \end{tikzpicture} |
|
685 \end{center} |
|
686 \begin{itemize} |
|
687 \item |
|
688 Lexer? Not just matcher |
|
689 \begin{itemize} |
|
690 \item |
|
691 Bit-Coded Regular Expression Parsing (Lasse Nielsen, Fritz Henglein 2011) |
|
692 \item |
|
693 Bit-Coded Regular Expression Parsing with Simplification (Sulzmann and Lu 2014) |
|
694 \end{itemize} |
|
695 |
|
696 \end{itemize} |
|
697 \end{frame} |
|
698 |
|
699 |
|
700 |
|
701 |
|
702 |
|
703 |
|
704 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
705 |
|
706 \begin{frame}[t] |
|
707 \frametitle{Ongoing Work} |
|
708 |
|
709 |
|
710 \begin{tikzpicture} |
|
711 \begin{semilogyaxis}[ |
|
712 xlabel={$n$}, |
|
713 x label style={at={(1.05,0.0)}}, |
|
714 ylabel={regex size}, |
|
715 enlargelimits=false, |
|
716 xtick={1,4,...,30}, |
|
717 xmax=33, |
|
718 %ytick={0, 10,...,100}, |
|
719 scaled ticks=false, |
|
720 axis lines=left, |
|
721 width=9cm, |
|
722 height=4.5cm, |
|
723 legend entries={Naive, Simp}, |
|
724 legend pos= north west, |
|
725 legend cell align=left |
|
726 ] |
|
727 \addplot[black,mark=square*, mark options={fill=white}] table {re-usize.data}; |
|
728 \addplot[black,mark=triangle*, mark options={fill=white}] table {re-ssize.data}; |
|
729 %\addplot[red,mark=triangle*,mark options={fill=white}] table {re1.data}; |
|
730 %\addplot[green,mark=square*,mark options={fill=white}] table {re2.data}; |
|
731 \end{semilogyaxis} |
|
732 \end{tikzpicture} |
|
733 |
|
734 \begin{itemize} |
|
735 \item |
|
736 Correctness of Sulzmann \& Lu's algorithm |
|
737 |
|
738 \item |
|
739 Size Bound $O(n^2t^2)$ of Dervatives (1995 Antimirov "Partial Derivative") |
|
740 \item |
|
741 More radical simplifications. |
|
742 \item |
|
743 Extend the algorithm to back-references. |
|
744 \end{itemize} |
|
745 \end{frame} |
|
746 |
|
747 |
|
748 |
|
749 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
|
750 |
|
751 |
|
752 |
|
753 |
|
754 |
|
755 |
|
756 |
899 |
757 |
900 |
758 |
901 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
759 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
902 \begin{frame}[c] |
760 \begin{frame}[c] |
903 \frametitle{\begin{tabular}{c}\\[3cm]\alert{Questions?}\end{tabular}} |
761 \frametitle{\begin{tabular}{c}\\[3cm]\alert{Questions?}\end{tabular}} |