On Thu, 28 Aug 2003 12:56:58 +0100 "Matt Brocklehurst" <Matt.Brocklehurst@xxxxxxxxxxxx> wrote: > Sorry for the two posts, didn_t notice your other reply! (Getting hungry > nearly lunch time!) > > Well im thinking if I fork the master process and use parent to do the > capturing continusoly and the child process to do the write, and using a > pipe between the two to get the data between the two. All theory atm but > im thinking it shuld work quite well. Just keep in mind that pipes are not asynchronous in Linux in applications started via popen(). So, if the client process wants to read from, say, stdin, that is OK. But you cannot use SIGIO on stdin. You will need to do explicit reads. Perhaps this is fine in your app. I did a similiar thing there I expected signals in the child process when I wrote to its stdin, but they did not arrive. Upon checking, I found that this was expected behavior in Linux. Lost a few hours there! > RE: JPEG compression - the libjpeg is screwed on my development box and I > really cant be bothered fixing it!!!! So is there any alternatives than > using this library, e.g. someone writen their own code? I really don_t > have the time/energy or brain! To do my own. Well not until I go on > another cafffeeeeene high. I use the stock IJPEG library for this (/usr/lib/libjpeg.so.62 on my system). You can have mine if you want. It is binary compatible with /lib/libc.so.6, which I bet your app is using if it is less than a few years old. 'ldd ./yourapp' to see. How is your library screwed? You can override using the system libjpeg on your development system while leaving the production using the system one, without modifying your binary (regression issues aside). It is all in the LD_RUN_PATH environment variable. -- +····························+·······························+ · Roger Oberholtzer · E-mail: roger@xxxxxx · · OPQ Systems AB · WWW: http://www.opq.se/ · · Erik Dahlbergsgatan 41-43 · Phone: Int + 46 8 314223 · · 115 34 Stockholm · Mobile: Int + 46 733 621657 · · Sweden · Fax: Int + 46 8 302602 · +····························+·······························+
