zipb

Argh! My small example patch awaits approval. Just over 40 KB. Let's 

see if this message comes trough.... Input is a list of chars with 

extra spaces, like z i p @ k n o w a r e . n l

First regexp get rids of the spaces, the second does a validity check 

Argh! My small example patch awaits approval. Just over 40 KB. Let's  
see if this message comes trough.... Input is a list of chars with  
extra spaces, like z i p @ k n o w a r e . n l

First regexp get rids of the spaces, the second does a validity check  
of some sorts...

￼
Zip Boterbloem
Media Mechanics
Zwaluwstraat 54
2025 VR  Haarlem
The Netherlands
+31627014758
zip@knoware.nl


regexp-troubles-expert-help-needed

On 9 nov. 06, at 17:12, Zip Boterbloem wrote:

> Argh! My small example patch awaits approval. Just over 40 KB. 

> Let's see if this message comes trough.... Input is a list of chars 

> with extra spaces, like z i p @ k n o w a r e . n l

> First regexp get rids of the spaces, the second does a validity 

Strange... it works here (PPC). I just put the - in the class of 

character at the beginning (it seems PCRE is more intelligent and 

knows when - have the special meaning inside a class). Can you 

confirm the crash with the following patch?

> Argh! My small example patch awaits approval. Just over 40 KB.  
> Let's see if this message comes trough.... Input is a list of chars  
> with extra spaces, like z i p @ k n o w a r e . n l
>
> First regexp get rids of the spaces, the second does a validity  
> check of some sorts...

Strange... it works here (PPC). I just put the - in the class of  
character at the beginning (it seems PCRE is more intelligent and  
knows when - have the special meaning inside a class). Can you  
confirm the crash with the following patch?

Thanks for having a look! Much appreciated.

It doesn't crash on any rogue(not a valid email address) input, but 

every once in a while(3 times in the last hour). Can't see a pattern 

yet. Lots of %@. characters seem to trigger it. I should save all 

input to a file, so I can see what causes it, but I'm just to busy 

for this sort of bug hunting. Anyway, methinks it shouldn't crash on 

Op 9-nov-2006, om 17:56 heeft Emmanuel Jourdan het volgende geschreven:

> On 9 nov. 06, at 17:12, Zip Boterbloem wrote:

>> Argh! My small example patch awaits approval. Just over 40 KB. 

>> Let's see if this message comes trough.... Input is a list of 

>> chars with extra spaces, like z i p @ k n o w a r e . n l

>> First regexp get rids of the spaces, the second does a validity 

> Strange... it works here (PPC). I just put the - in the class of 

> character at the beginning (it seems PCRE is more intelligent and 

> knows when - have the special meaning inside a class). Can you 

> confirm the crash with the following patch?

> #P hidden newex 411 150 219 196617 loadmess set;

> #P message 35 67 64 196617 to- to @ti ti;

> #P message 140 186 119 196617 to-to@titi;

> #P newex 140 157 62 196617 prepend set;

> #P message 304 281 119 196617 to-to@titi;

> #P newex 304 252 62 196617 prepend set;

> #P message 241 338 119 196617 zip@knoware.nl;

> #P newex 241 309 62 196617 prepend set;

> #P newex 118 130 188 196617 regexp @re (\\S)\\s? @substitute %1;

> #P message 159 93 119 196617 z i p @ k n o w a r e . n l;

> #P message 118 67 82 196617 to -to@ti ti.com;

> #P newex 118 218 219 196617 regexp [-a-z0-9._%]+@[-a-z0-9.]+\\.[a- 

It doesn't crash on any rogue(not a valid email address) input, but  
every once in a while(3 times in the last hour). Can't see a pattern  
yet. Lots of %@. characters seem to trigger it. I should save all  
input to a file, so I can see what causes it, but I'm just to busy  
for this sort of bug hunting. Anyway, methinks it shouldn't crash on  
any input, overdrive on or not.

Zip Boterbloem
Media Mechanics
Zwaluwstraat 54
2025 VR  Haarlem
The Netherlands
+31627014758
zip@knoware.nl

> On 9 nov. 06, at 17:12, Zip Boterbloem wrote:
>
>> Argh! My small example patch awaits approval. Just over 40 KB.  
>> Let's see if this message comes trough.... Input is a list of  
>> chars with extra spaces, like z i p @ k n o w a r e . n l
>>
>> First regexp get rids of the spaces, the second does a validity  
>> check of some sorts...
>
> Strange... it works here (PPC). I just put the - in the class of  
> character at the beginning (it seems PCRE is more intelligent and  
> knows when - have the special meaning inside a class). Can you  
> confirm the crash with the following patch?
>
> ej
>
> #P window setfont "Sans Serif" 9.;
> #P window linecount 1;
> #P hidden newex 411 150 219 196617 loadmess set;
> #P message 35 67 64 196617 to- to @ti ti;
> #P message 140 186 119 196617 to-to@titi;
> #P newex 140 157 62 196617 prepend set;
> #P message 304 281 119 196617 to-to@titi;
> #P newex 304 252 62 196617 prepend set;
> #P message 241 338 119 196617 zip@knoware.nl;
> #P newex 241 309 62 196617 prepend set;
> #P newex 118 130 188 196617 regexp @re (\\S)\\s? @substitute %1;
> #P message 159 93 119 196617 z i p @ k n o w a r e . n l;
> #P button 222 307 15 0;
> #P button 274 281 15 0;
> #P message 118 67 82 196617 to -to@ti ti.com;
> #P newex 118 218 219 196617 regexp [-a-z0-9._%]+@[-a-z0-9.]+\\.[a- 
> z]+;
> #P connect 4 0 5 0;
> #P connect 1 0 5 0;
> #P connect 12 0 5 0;
> #P connect 5 0 0 0;
> #P connect 5 0 10 0;
> #P hidden connect 13 0 11 0;
> #P connect 10 0 11 0;
> #P connect 0 2 3 0;
> #P connect 0 2 6 0;
> #P hidden connect 13 0 7 0;
> #P connect 6 0 7 0;
> #P connect 0 3 2 0;
> #P connect 0 3 8 0;
> #P hidden connect 13 0 9 0;
> #P connect 8 0 9 0;
> #P window clipboard copycount 14;
>
>


I agree, but unless you're willing to invest the time to show how it 

crashes, It can't be fixed. Also, please, in the future, send 

Am 09.11.2006 um 18:15 schrieb Zip Boterbloem:

> Anyway, methinks it shouldn't crash on any input, overdrive on or not.

I agree, but unless you're willing to invest the time to show how it  
crashes, It can't be fixed. Also, please, in the future, send  
patches, not pictures.

> Anyway, methinks it shouldn't crash on any input, overdrive on or not.


I 've sent patches. But they're stopped by the moderator(they are 

just over 40 k), who doesn't send them on. So this was the best I 

Op 10-nov-2006, om 14:10 heeft Jeremy Bernstein het volgende geschreven:

> I agree, but unless you're willing to invest the time to show how 

> it crashes, It can't be fixed. Also, please, in the future, send 

> Am 09.11.2006 um 18:15 schrieb Zip Boterbloem:

>> Anyway, methinks it shouldn't crash on any input, overdrive on or 

I 've sent patches. But they're stopped by the moderator(they are  
just over 40 k), who doesn't send them on. So this was the best I  
could do.

> I agree, but unless you're willing to invest the time to show how  
> it crashes, It can't be fixed. Also, please, in the future, send  
> patches, not pictures.
>
> jb
>
> Am 09.11.2006 um 18:15 schrieb Zip Boterbloem:
>
>> Anyway, methinks it shouldn't crash on any input, overdrive on or  
>> not.
>


Zip - we'd generally prefer if you send _small_ patches which clearly 

and obviously demonstrate the problem, without distractions like 

network communication, etc. Basically, a regexp object with a string 

that breaks it is all we need. And that won't be 40k.

Am 10.11.2006 um 15:37 schrieb Zip Boterbloem:

> I 've sent patches. But they're stopped by the moderator(they are 

> just over 40 k), who doesn't send them on. So this was the best I 

Zip - we'd generally prefer if you send _small_ patches which clearly  
and obviously demonstrate the problem, without distractions like  
network communication, etc. Basically, a regexp object with a string  
that breaks it is all we need. And that won't be 40k.

> I 've sent patches. But they're stopped by the moderator(they are  
> just over 40 k), who doesn't send them on. So this was the best I  
> could do.


On 10 nov. 06, at 15:37, Zip Boterbloem wrote:

You can also send part of the patch by just select/copy and paste the 

> Hi Jeremy,
>
> I 've sent patches. But they're stopped by the moderator(they are  
> just over 40 k), who doesn't send them on. So this was the best I  
> could do.

You can also send part of the patch by just select/copy and paste the  
contents as text (since MaxMSP 4.5.5).

I type some random characters, lots of %@. characters. 99 out of 100 

times this works OK. It crashes 1 out of 100 times, Max disappears 

taking the culprit string with it. Yesterday afternoon it crashed 

three times in an hour of testing. This is not good enough for 

unattended use in a museum installation. I don't have the time to log 

everything to a file. Must get big patch working before Monday.... 

Wiil revert to JS regular expressions for now.

Op 10-nov-2006, om 15:45 heeft Jeremy Bernstein het volgende geschreven:

> Zip - we'd generally prefer if you send _small_ patches which 

> clearly and obviously demonstrate the problem, without distractions 

> like network communication, etc. Basically, a regexp object with a 

> string that breaks it is all we need. And that won't be 40k.

> Am 10.11.2006 um 15:37 schrieb Zip Boterbloem:

>> I 've sent patches. But they're stopped by the moderator(they are 

>> just over 40 k), who doesn't send them on. So this was the best I 

I type some random characters, lots of %@. characters. 99 out of 100  
times this works OK. It crashes 1 out of 100 times, Max disappears  
taking the culprit string with it. Yesterday afternoon it crashed  
three times in an hour of testing. This is not good enough for  
unattended use in a museum installation. I don't have the time to log  
everything to a file. Must get big patch working before Monday....  
Wiil revert to JS regular expressions for now.

> Zip - we'd generally prefer if you send _small_ patches which  
> clearly and obviously demonstrate the problem, without distractions  
> like network communication, etc. Basically, a regexp object with a  
> string that breaks it is all we need. And that won't be 40k.
>
> jb
>
> Am 10.11.2006 um 15:37 schrieb Zip Boterbloem:
>
>> I 've sent patches. But they're stopped by the moderator(they are  
>> just over 40 k), who doesn't send them on. So this was the best I  
>> could do.
>


OK, I hope you'll take the time after your deadline to follow up.

Am 10.11.2006 um 16:59 schrieb Zip Boterbloem:

> I type some random characters, lots of %@. characters. 99 out of 

> 100 times this works OK. It crashes 1 out of 100 times, Max 

> disappears taking the culprit string with it. Yesterday afternoon 

> it crashed three times in an hour of testing. This is not good 

> enough for unattended use in a museum installation. I don't have 

> the time to log everything to a file. Must get big patch working 

> before Monday.... Wiil revert to JS regular expressions for now.

> I type some random characters, lots of %@. characters. 99 out of  
> 100 times this works OK. It crashes 1 out of 100 times, Max  
> disappears taking the culprit string with it. Yesterday afternoon  
> it crashed three times in an hour of testing. This is not good  
> enough for unattended use in a museum installation. I don't have  
> the time to log everything to a file. Must get big patch working  
> before Monday.... Wiil revert to JS regular expressions for now.
>
> Thanks for your help anyway.


I have a whole list of rogue objects/behaviour. If I ever get the 

Op 10-nov-2006, om 17:23 heeft Jeremy Bernstein het volgende geschreven:

> OK, I hope you'll take the time after your deadline to follow up.

> Am 10.11.2006 um 16:59 schrieb Zip Boterbloem:

>> I type some random characters, lots of %@. characters. 99 out of 

>> 100 times this works OK. It crashes 1 out of 100 times, Max 

>> disappears taking the culprit string with it. Yesterday afternoon 

>> it crashed three times in an hour of testing. This is not good 

>> enough for unattended use in a museum installation. I don't have 

>> the time to log everything to a file. Must get big patch working 

>> before Monday.... Wiil revert to JS regular expressions for now.

I have a whole list of rogue objects/behaviour. If I ever get the  
time...

> OK, I hope you'll take the time after your deadline to follow up.
>
> jb
>
> Am 10.11.2006 um 16:59 schrieb Zip Boterbloem:
>
>> I type some random characters, lots of %@. characters. 99 out of  
>> 100 times this works OK. It crashes 1 out of 100 times, Max  
>> disappears taking the culprit string with it. Yesterday afternoon  
>> it crashed three times in an hour of testing. This is not good  
>> enough for unattended use in a museum installation. I don't have  
>> the time to log everything to a file. Must get big patch working  
>> before Monday.... Wiil revert to JS regular expressions for now.
>>
>> Thanks for your help anyway.
>


regexp troubles, expert help needed