Download raw (9.1 KB)
<snelting> hello from brussels! <Tsoots> lo <mrdocs> hi femke <pierremarc> hi <snelting> pierre, you asked about the tiffs we are loading -- it is one at a time and we tried first with jpgs and this doesn't make a difference for speed. <pierremarc> snelting: It’s just that you load them in a loop & I don’t know if they are actually "unloaded" at some point <pierremarc> what a TOP gives? |<-- ludo has left freenode (Read error: 104 (Connection reset by peer)) <snelting> ah. (i'm a python absolute beginner) <snelting> 6336 snelting 25 0 398m 156m 13m R 99 16.1 593:33.01 scribus -->| sjc (n=sjc@78.32.104.102) has joined #scribus <snelting> wait... <snelting> that's 99% cpu for scribus ;-) <pierremarc> normal <pierremarc> It works after all ;) <snelting> yes, for almost 12 hours! <snelting> (still running ... at number 310 now. impressing) <snelting> impressive i meant * sjc drops in to follow the "Testing Scribus Limits" discussion -->| ludo (n=ludo@118.10.101-84.rev.gaoland.net) has joined #scribus <sjc> hi snelting <snelting> hello! <pierremarc> From what you wrote, I understand tyhat scribus generates some garbage somewhere which it in turn takes long time to process, but where? its still a question! <snelting> I didn't say garbage! <snelting> ;-) |<-- Okona has left freenode (leguin.freenode.net irc.freenode.net) |<-- Wed has left freenode (leguin.freenode.net irc.freenode.net) |<-- ben_goodger has left freenode (leguin.freenode.net irc.freenode.net) <snelting> it is still relatiely fast, but with 200 more runs to go... what do you think? will scribus make it to the end? -->| ben_goodger (n=ben@host86-146-18-220.range86-146.btcentralplus.com) has joined #scribus -->| Wed (n=Wed@83.209.33.104) has joined #scribus -->| Okona (n=Okona@macrbg1.informatik.tu-muenchen.de) has joined #scribus <snelting> i mean... i think we have a bit of extra time so if it means waiting than we will just be patient <pierremarc> cool <pierremarc> Although there’s something that looks like a memory leak here <snelting> but it would be tragic if the process stops at 498... yes, something is leaking or clogging up <sjc> has anything been concluded in the code yet before I arrived? I did not like those LoadImage lines when I quickly glanced through the source <pierremarc> Oh, for 2 covers, you can trick it by making them by hand :D <snelting> good thinking ... duh <pierremarc> Printer deadline postoned is the fix for now sjc! <sjc> pierremarc: oic :) <snelting> so i should look for some kind of 'unloading' in the script? is scribus memorizing all former 298 docs you think? that would be really amazing! <pierremarc> I don’t think so <pierremarc> I’ll try to reproduce it to see <snelting> ok, i can upload the folder structure with the first 5 elements plus .sla? <pierremarc> Oh, it would be great <snelting> ok give me a few minutes <snelting> it is great to have you look at it -- by now getting really curious about what is going on -->| Asynchronicle (n=Asynchro@68.244.193.82.ediscom.de) has joined #scribus <snelting> ok, it should be ok -- i don't dare to launch another scribus to test right now... <pierremarc> I will :) <snelting> http://ospublish.constantvzw.org/sources/pzicatalogue <mrscribe> Title: Index of /sources/pzicatalogue (at ospublish.constantvzw.org) <snelting> so now you should be able to generate up to 5 catalogues <snelting> (at cover no. 324 now) <pierremarc> snelting: sorry to bother you but, can you zip or tar or whatever to get it all in only one archive? It will ease the download a lot <snelting> ok you are right <snelting> sorry! <snelting> there! http://ospublish.constantvzw.org/sources/pzicatalogue/cataloguemachine.zip <pierremarc> c’est parti !!! -->| Cymek (n=Cymek__@p4FF65D06.dip.t-dialin.net) has joined #scribus <snelting> tell me if you are missing anything? <pierremarc> it complains about missing images, but I"ll modify it <snelting> ok, shouldn't make a difference to leakage or clogging ;-) |<-- Asynchronicle has left freenode ("left") |<-- Herm has left freenode (Read error: 110 (Connection timed out)) <pierremarc> snelting: checking the script & material to change things I see that gordo’s images are in fact PDFs, why? <snelting> hmmm. good question. all material comes from different art projects; this is how he generated it and we have left it that way after testing with the printer. <snelting> not elegant, i admit <snelting> he = gordo <pierremarc> ok <snelting> also, we were interested in working with many different source materials -- each of the graduates in the catalogue has found a way to add something to each of the 500 covers <snelting> the 'lettriste' you see in the python script is one of them, randomly changing letter sizes <pierremarc> well, it’s cool but if it ends up to be an issue, might be interesting to first convert all to an unique file type, no? * pierremarc is still so slow at typing :) <snelting> yes. if we need to we will do it of course. <snelting> the placed pdf is a bit silly <snelting> especially because it is three giant pixels (if you put all catalogues together, it forms an image) <snelting> scribus still holding out well... at no. 332 now <--| JanH has left #scribus -->| JanH (n=joerg@Pb614.p.pppool.de) has joined #scribus <--| JanH has left #scribus <snelting> pierre: ça va? <snelting> i will need to go out for a few hours <pierremarc> I monitor the script now <snelting> leaving the catalogue machine running on its own <snelting> how do you monitor? <pierremarc> Right now just with timers. Want to reproduce your issur first <snelting> ok. i'm just looking at the difference of the writing time of the files <pierremarc> Images(0) loaded in 1.46746706963 <pierremarc> PDF(0) saved in 22.6262660027 <pierremarc> Images(1) loaded in 1.37868094444 <pierremarc> PDF(1) saved in 22.6658248901 <pierremarc> Images(2) loaded in 1.39984512329 <pierremarc> PDF(2) saved in 21.3724861145 <pierremarc> Images(3) loaded in 1.52945899963 <pierremarc> PDF(3) saved in 22.4700379372 <pierremarc> Images(4) loaded in 1.68279314041 <pierremarc> PDF(4) saved in 24.7134678364 <snelting> so... this means between 22 and 24 secs? <pierremarc> yes <snelting> hmm so it does go up, but don't know what you can tell from 5 runs <snelting> sorry to have to leave now in the middle of the detective <pierremarc> np <pierremarc> I will be here at least tonight <snelting> ok, me too <snelting> thanks for a first look! <pierremarc> de rien, le projet a l’air sympa LATER.... <snelting> good news -- just finished pdf nr. 500 <pierremarc> nice <a_l_e> ... <snelting> i mean; so scribus has worked for 23 hours non-stop <pierremarc> really, cause i was going to bed :) <snelting> hihi. so interesting is, that it did not slow down much more <snelting> i will go through the script with some people in the coming days to see what can be done better <snelting> (for the next time ;-)) <MrB> were the same fonts used for all pages? <snelting> yes. <MrB> ie, for all docs, or different ones for some of the front covers/ <MrB> ? <snelting> the inside is identical for each book; the cover has our own notcourier-sans only <snelting> inside is manually generated <MrB> k <snelting> we had problems earlier, because of a corrupt font. <snelting> 'manually' generated pages (from scribus, selecting 'export as pdf' manually) |<-- jghali has left freenode ("Parti") <MrB> yeah <snelting> would automatically detect the problem and outline it -->| jghali (n=jghali@ANantes-157-1-159-93.w86-195.abo.wanadoo.fr) has joined #scribus =-= Mode #scribus +o jghali by mrscribe <snelting> but the python generated pdf embedded it / tried to deal with it and that made huge files that made the printers pre-flight checker crassh! -->| hawk_pdm (n=robert@dslb-088-072-238-036.pools.arcor-ip.net) has joined #scribus <snelting> so, that was a good lesson. <MrB> it shouldnt have been any different in theory <MrB> hm <MrB> luckily its getting a rewrite with GSoC <snelting> i am looking forward! combined with POD the scripter really makes a difference -->| JanH (n=joerg@Pb614.p.pppool.de) has joined #scribus <snelting> our experiments are a bit brutal now <MrB> hehe <snelting> but we need to start somewhere ;-) <snelting> ok, i'll go over the script with some people and post back about what we find on our side <MrB> ok <MrB> thanks <mrdocs> snelting: hi femke <snelting> hello! <mrdocs> btw, it looks like amsterdam is the leading candidate for LGM so far <snelting> yes i saw <snelting> if it is going to go through, there is many people i should connect lgm to <snelting> it is a bit lazy option for me / us but it could work out well if lgm lands in the land of design ;-) <snelting> i am not a good organiser but i will do what i can to help <mrdocs> :) <snelting> if it is final, amsterdam it will be on the create list? <MrB> the create list is used for lgm organisation, yes, so you will see it there <snelting> ok. i'm already looking out for things -- for sure the piet zwart institute (the org we did the 500 cover catalogue for) can be helpful <snelting> i'll gather my notes <snelting> thanks again, and pierre: sleep well!