| ID |
Date |
Icon |
Author |
Author Email |
Category |
OS |
ELOG Version |
Subject |
|
1761
|
Tue Mar 7 21:41:28 2006 |
 | Haitao Yu | htyu@phys.columbia.edu | Question | Linux | 2.6.1 | large attachment seems to hang elogd? |
Hello,
I am trying to attach somewhat large file (4-15MB files tested) with my messages but it seems to hang the elogd server for a really long time (30-100minutes). The file upload was done pretty soon (within a few seconds), then the elogd process seems to be busy doing something, using around 50MB of memory and 95% of CPU time, and stop respond to the requests. I can read the logbook with another copy of elogd running on another port, and I can see (and download) the uploaded file. I am wondering what could cause the problem --- and could it be related to the problem of not setting up the SMTP server right? Thanks for any suggestion!
Haitao
P.S. I read in some old message that if we want to upload very large files, we have to change the WEB_BUFFER_SIZE in elogd.c, however this parameter is no longer in v2.6.1. Is there something similiar I have to change? |
|
1760
|
Mon Mar 6 20:51:57 2006 |
 | Glenn Horton-Smith | gahs@phys.ksu.edu | Request | All | 2.6.1-1668 | Provide option to require smileys to be bracketed a la ELCode? |
| Stefan Ritt wrote: |
| Glenn Horton-Smith wrote: | It would be nice if elog would only interpret something as a smiley if it is surrounded by whitespace. It can be particularly annoying that an 8 followed by a right paren becomes a "cool" smiley -- e.g., a parenthetical reference to event eighteen (18) becomes mangled... [That was "18" inside the parens.] ...
|
Interpreting smileys only if they are surrounded by whitespace does not solve the problem completely. It will solve it for (1\8), but not if you have (1, \8) (1, 9) in your text. So it's not a good solution. If you have problems with simleys, I would post my text in plain mode, or surround your numbers with [code]...[/code] tags...
|
Hmm, you're right. But I like EL Code, and I don't want to give it up just because of the smileys! A better solution would be to provide an option, which when set, would require smileys to be in brackets, e.g., [\8)] would become the cool smiley [8)], rendered without surrounding brackets if the option was set.
I actually just spent some time I didn't really have and modified my copy of elogd to see if it could be done and would work as intended. Eureka, it works! The modification implements an option "use bracketed smileys" which, if set to 1, causes the editor smiley bar to insert smileys with the square brackets around them and causes rsputs_elcode() to only substitute for smileys if there are square brackets around them, suppressing the square brackets.
I like this feature. Maybe others would like it too. I have a patch file (diff -cbw) w.r.t. svn version 1663 if you're interested. |
|
1759
|
Mon Mar 6 18:54:32 2006 |
 | Steve Jones | steve.jones@freescale.com | Bug report | All | 2.6.1-1660 | [UPDATE] Re: Crafted URL causes elog to coredump |
| Steve Jones wrote: |
| Stefan Ritt wrote: |
No, the above link just works fine, just click it. |
| Quote: | I was afraid to try  . Ok, then the issue *might* be rev 1660. On my production version running 2.5.3 I get the expected "Invalid URL: <name>" box. Any suggestions on how to find out? |
|
| Steve Jones wrote: | Ok, here is what I found:
- Now running rev 1671
- Specifying a non-existent logbook with plain URL of elog server: OK, error is caught and reported
- Specifying a non-existent logbook *under* a logbook name: Ok, error is caught.
- Specifying a non-existent logbook under a "Top Group" of the form "This+Is+Top+Group+One" results in no error, elogd crash. Example URLs:
- Results in a program crash under Solaris. Top Groups were not defined under elog 2.5.3.
|
|
|
1758
|
Mon Mar 6 18:36:34 2006 |
 | Steve Jones | steve.jones@freescale.com | Question | All | 2.5.9 | [RESOLVED] eLog Version number as eLog attribute? |
| Steve Jones wrote: |
| Stefan Ritt wrote: |
| Steve Jones wrote: | When a footer is used (via Bottom text = <filename>) eLog no longer displays the eLog version number at the bottom. Is it possible to somehow expose the version/revision as an eLog attribute or have the version still display even when a replacement footer is specified?
Thanks! |
I added that feature, but will not be able to commit it before the next weekend. |
| Quote: | | Not a problem! Thanks |
|
| Steve Jones wrote: | | Works as requested!! |
|
|
1757
|
Mon Mar 6 18:06:32 2006 |
| Steve Jones | steve.jones@freescale.com | Bug report | All | 2.6.1-1660 | Re: Crafted URL causes elog to coredump |
| Stefan Ritt wrote: |
No, the above link just works fine, just click it. |
| Quote: | | I was afraid to try . Ok, then the issue *might* be rev 1660. On my production version running 2.5.3 I get the expected "Invalid URL: <name>" box. Any suggestions on how to find out? |
|
|
1756
|
Mon Mar 6 18:04:39 2006 |
| Steve Jones | steve.jones@freescale.com | Bug report | All | 2.6.1-1660 | Re: Crafted URL causes elog to coredump |
| Stefan Ritt wrote: |
No, the above link just works fine, just click it. |
| Quote: | | I was afraid to try . Ok, then the issue *might* be rev 1660 or perhaps the fact that compiled under Solaris it does this. Any suggestions on how to find out? |
|
|
1755
|
Mon Mar 6 17:45:18 2006 |
| Stefan Ritt | stefan.ritt@psi.ch | Bug report | All | 2.6.1-1660 | Re: Crafted URL causes elog to coredump |
No, the above link just works fine, just click it. |
|
1754
|
Mon Mar 6 17:43:11 2006 |
| Steve Jones | steve.jones@freescale.com | Request | All | 2.6.1 | Re: Allow $attributes in "Comment = " option |
| Stefan Ritt wrote: |
| Steve Jones wrote: | | Is it possible to allow $attribute substitutions in the "Comment =" option for logbooks? |
The "Comment =" option for logbooks gives a general comment, like Discussion forum about ELOG for this forum. Since this comment is global, it does not make sense to have $attribute substitution. |
| Quote: | | Ok, just a thought. Thanks |
|