Aidan> SuSE reported problems with pdump on IA 64, but I'd try compiling Aidan> --with-pdump before being confident that XEmacs won't work there. IIRC the IA-64's architecture is significantly incompatible with pdump. This is because function pointers are really function descriptors that can contain a field that is not guaranteed to be constant across invocations. XEmacs will build and even pass the unit tests but might not work after the machine is rebooted. SuSe had a pretty invasive patch to fix this (by reinitialised all function pointers in dumped data upon startup) but it's never made it into our official releases. Malcolm - Typing this from just rebuilt Linux PPC64 partition. -- Malcolm Purvis <malcolmp(a)xemacs.org&gt;

Aidan> That's an aspect of the ABI, no? And may vary on HP-UX? No, I think that the ABI is the same at this point. I've found links to some more details about this problem. The last time it came up in -beta was in Feburary: http://thread.gmane.org/gmane.emacs.xemacs.beta/17972 I still have the dusty workspace with a half-cleaned up patch. I haven't tried to submit it because it is quite invasive and only affects a very small number of users. If there is more interest then I'll get it ready for submission. A more detailed description of the problem can be found at: http://article.gmane.org/gmane.lisp.gcl.devel/2444 Malcolm -- Malcolm Purvis <malcolmp(a)xemacs.org&gt;

Albert> Does CANNOT_DUMP work? I've seen it in the code but this is the first time I've ever seen it mentioned by a human being. Albert> If not, anyone interested in a contract to get XEmacs Albert> unexec working (and possibly for Emacs 21.x as well)? Maybe I should let somebody take your money before asking :-) but have you tried the "portable dumper" solution (./configure --pdump [your favorite options here])? -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

Ar an naoiú lá de mí Eanair, scríobh Darryl Okahata: My experience is consistent with this; I’m using this XEmacs (21.4) with a 46M RSS size limit. I load VM and ERC and open mail folders serially, with my inbox being 4.8M on disk. At some point a malloc necessary to load a mail folder into a buffer will eventually fail; even though I don’t load any more packages and am scrupulous about killing unused buffers, this is unavoidable. There is a effective memory leak somewhere. I’m using the relocating allocator too. -- I AM IN JAIL AND ALLOWED SEND ONLY ONE CABLE SINCE WAS ARRESTED WHILE MEASURING FIFTEEN FOOT WALL OUTSIDE PALACE AND HAVE JUST FINISHED COUNTING THIRTY EIGHT THOUSAND FIVE HUNDERED TWENTY TWO NAMES WHOS WHO IN MIDEAST.

Ar an t-aonú lá déag de mí Eanair, scríobh Stephen J. Turnbull: I am, that’s what reading mailboxes serially involves. I’m seeing this effective leakage on 21.4.17 with just TTY support--that is, no X libraries, no Xft libraries. LDAP and TLS support, though, for some reason. I don’t trust the memory usage stats in 21.5. I routinely have resident set sizes of 500 MB on my desktop, and the total memory usage XEmacs tells me about is a ridiculously tiny fraction of that. Is anyone else seeing the incremental garbage collector tie up their process at least as much as the old one? And is the reason no-one else has posted a problem report on that crash of mine the other day because everyone’s turned off incremental GC? Because I got the same crash again yesterday evening, so it doesn’t appear to be anything atypical in my configuration. -- I AM IN JAIL AND ALLOWED SEND ONLY ONE CABLE SINCE WAS ARRESTED WHILE MEASURING FIFTEEN FOOT WALL OUTSIDE PALACE AND HAVE JUST FINISHED COUNTING THIRTY EIGHT THOUSAND FIVE HUNDERED TWENTY TWO NAMES WHOS WHO IN MIDEAST.

Aidan> I am, that’s what reading mailboxes serially involves. In what mailreader? Also, I don't know the relocating allocator, does it try to return memory every time a large buffer is deallocated, or only if memory gets to be some fraction less than peak? Aidan> I’m seeing this effective leakage on 21.4.17 with just TTY Aidan> support--that is, no X libraries, no Xft libraries. LDAP Aidan> and TLS support, though, for some reason. OK. Aidan> I don’t trust the memory usage stats in 21.5. I routinely Aidan> have resident set sizes of 500 MB on my desktop, and the Aidan> total memory usage XEmacs tells me about is a ridiculously Aidan> tiny fraction of that. Hm? I routinely run for weeks, and usually equilibrate at about 150MB. As an XEmacs developer I'm not proud of that, but 500MB I've never seen, and it's been a year or so since I've seen 200MB. Aidan> Is anyone else seeing the incremental garbage collector tie Aidan> up their process at least as much as the old one? Heh. I'm not going to touch the IGC until I get my new box running, I've got too much time wasted dealing with disk space at 95% or more on all partitions.... -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

Aidan> Yup. (With my later caveat that it’s virtual memory, not Aidan> RSS, of 500MB.) Steve Youngs mentions on the SXEmacs lists Aidan> recently a process of his that broke one gigabyte, Good grief. We need to find out where all that is going, for sure. Aidan> I’m not astonished at this, but I do get the impression Aidan> things are being leaked. I am. Gnus doesn't keep that much around in news (unless you're reading binaries), where "that much" is measured in MB, maybe 5-10 is as high as I can imagine. Your VM is gonna be similar, right? I do edict (Japanese dictionary, whole thing in memory) so I can see the En-De going 10MB or so, Quail (except for CJK) is small (though Mule-UCS is not) but even so 1, 2 MB. I see 50MB in buffers, tops. Where's the other 450MB going? I wonder if lstreams leak. Aidan> I’m hacking too much C lately to leave my desktop XEmacs Aidan> running for weeks; I know the feeling, although with 1GB on the Mac I can run 2-3 XEmacs processes, and most of my C hacking recently has been modules so I can usually reload a couple of times before things blow up. :-) Aidan> and my remove-server XEmacs YOW! Now THAT is quite a BUG, we Japanese would like to use IT on th' local Axis-of-Evil DICTATOR! (Even rounding the v to u doesn't help. ;) -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

Aidan> MOUE, n; a pouting expression, often conveying (mock) Aidan> annoyance or distaste, or used flirtatiously. Aidan> REMOUE-SERVER, a; used to describe a process causing Aidan> annoyance or distaste to a server. A _flirtatious_ server. Batting its eyes at Brat-packer script kiddies, no doubt. -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.