On Sat, Apr 01, 2006 at 08:55:30AM -0400, Bruno Barrera C. wrote: > On Sat, 2006-04-01 at 01:49 -0500, Justin Pryzby wrote: > > tag 359854 patch > > thanks > > > > It is looping with recv()=0, which means "the remote end has shut > > down". It is a special return value, and has to be handled as such. > > Attached is functional and mildly tested patch. > > > > I also made some changes to fix some ugly stuff valgrind turned up. > > > > It still doesn't detect errors, but this didn't work before anyway. > > > > I would seriously reconsider maintenance of this package..trivially > > fixed valgrind warnings, saved files aren't correct, overwrites files > > which wget wouldn't, and improper use of recv. > > Could you please explain me what does "reconsider maintenance of this > package" means? I am being critical of the state of the software, wondering if it shouldn't be removed from the archive in favor of other downloader tools (which I haven't looked into closely).
> Are referering to my work or upstream?. I noticed the large .diff.gz :) Although I note that the unpatched aget doesn't loop in recv, as the patched one does. I'm not sure what can be the cause of this; I have read the .diff.gz, and my comments follow. My best guess now is that the thread/signal changes are to blame. Aget.c: @@ -98,13 +87,12 @@ @@ -176,32 +154,22 @@ Head.c: @@ -74,7 +72,7 @@ snprintf change is backwards? Download.c @@ -86,14 +81,19 @@ Loop exploitable, can crash client double isn't appropriate for filesizes; it is actually less accurate than an unsigned long long int or off_t or whatever. Compare ~15 significant figures with ~19. But I don't think this is the problem, since the PDF is <<1GB. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
