calibre 0.6 betas now available

[sirbruce]

Originally Posted by Gideon

Anyone else seeing that Calibre freezes up after adding a few PDFs? I get about two in and then on the third it just stays on the adding window perpetually until I abort.

I'm seeing it now, Gideon.

1. There are some PDFs Calibre won't read at all; it'll just sit there stuck when it gets to one. [Update: Actually, it appears it'll read those troublesome PDFs individually, but not if imported in a batch with other PDFs.)

2. When it does read a PDF, it has to read in and parse the entire file; takes a couple of minutes.

3. After adding more than a few PDFs at a time, Calibre's process size will balloon to about 2GB. My computer then starts paging it and peformance goes into the toilet. Aborting the import frees the memory and you can re-import again normally until it happens again.

[user_none]

Originally Posted by sirbruce

I'm seeing it now, Gideon.

1. There are some PDFs Calibre won't read at all; it'll just sit there stuck when it gets to one. [Update: Actually, it appears it'll read those troublesome PDFs individually, but not if imported in a batch with other PDFs.)

2. When it does read a PDF, it has to read in and parse the entire file; takes a couple of minutes.

3. After adding more than a few PDFs at a time, Calibre's process size will balloon to about 2GB. My computer then starts paging it and peformance goes into the toilet. Aborting the import frees the memory and you can re-import again normally until it happens again.

Can you open a ticket and attach a few of the PDFs that display this behavior.

Also, can you try adding them with the PDF metadata reader plugin disabled and report back on how it works then.

[sirbruce]

Well I created a ticket about PDF import being slow and kovid closed it with the implication that PDF support is going to suck and there's little to be done about it. I was just converting a batch of PDFs and somehow it consumed all available memory on my Windows XP box, even though no process size showed the memory being consumed, and I had to reboot. I don't anticipate I'm going to be trying bulk converts of PDFs again.

I could zip up a collection and give them to you to play with but I'm reluctant to waste more time on it as getting the LIT and other bugs I've filed fixed are more important to me.

[user_none]

Originally Posted by sirbruce

I could zip up a collection and give them to you to play with but I'm reluctant to waste more time on it as getting the LIT and other bugs I've filed fixed are more important to me.

Send them to me. At the very least I can see if it's a Calibre issue or something to do with the one of the PDF libraries Calibre uses for things like the metadata.

[sirbruce]

Originally Posted by user_none

Send them to me. At the very least I can see if it's a Calibre issue or something to do with the one of the PDF libraries Calibre uses for things like the metadata.

I'll upload them to MegaUpload (they're big). Things to try:

1. Try importing them all at once, wait for a hang.

2. Try importing them 10 at a time, watch Calibre's process size.

3. Try importing them 1 at a time (usually works).

4. Try converting them all at once, watch your computer grind to a halt.

[elakin]

When uploading PDFs to a kindle DX with 0.6b6, it first converts 'em. is there something i have to do to have the pdf's uploaded as-is?

[sirbruce]

Originally Posted by elakin

When uploading PDFs to a kindle DX with 0.6b6, it first converts 'em. is there something i have to do to have the pdf's uploaded as-is?

Click on the downarrow next to "Send to Device" and select the option to send a specific format to the device.

A future patch is supposed to send PDFs to the DX automatically.

[sirbruce]

Well I've dumped 4,000 books through Calibre 0.6b6 so far, and filed over 20 bug reports. Right now I'm only converting stuff to MOBI; I assume most of the bugs I find will occur when converting to other formats as well, so I haven't been testing those, but I can.

I've got a few PDB I can test as well.

[user_none]

Originally Posted by elakin

When uploading PDFs to a kindle DX with 0.6b6, it first converts 'em. is there something i have to do to have the pdf's uploaded as-is?

PDF was added as a supported format for the DX after beta 6 was released. It will work correctly with the next release.

[user_none]

Originally Posted by sirbruce

I'll upload them to MegaUpload (they're big). Things to try:

1. Try importing them all at once, wait for a hang.

2. Try importing them 10 at a time, watch Calibre's process size.

3. Try importing them 1 at a time (usually works).

4. Try converting them all at once, watch your computer grind to a halt.

I can't get it to hang. It shows the adding progress bar and adds about 1 per second. I tried 6 times, same result. I also watched the process size each time and it remains at a consistent and reasonable amount.

I converted them all at once. There was a slight hang in the app while it "thought" about what needed to be done (I'll look into alleviating this if possible). The conversion went smoothly. Two books at a time (one per processor core). The computer stayed responsive. I'm assuming you were converting to Mobi? It looks like in some cases there is a memory issue with mobi so that is probably why you were experiencing this.

0.6 has a new jobs system and like Kovid, I use Linux. What you might be seeing with adding might be a Windows specific bug in the new code.

[sirbruce]

Originally Posted by user_none

I can't get it to hang. It shows the adding progress bar and adds about 1 per second. I tried 6 times, same result. I also watched the process size each time and it remains at a consistent and reasonable amount.

I converted them all at once. There was a slight hang in the app while it "thought" about what needed to be done (I'll look into alleviating this if possible). The conversion went smoothly. Two books at a time (one per processor core). The computer stayed responsive. I'm assuming you were converting to Mobi? It looks like in some cases there is a memory issue with mobi so that is probably why you were experiencing this.

0.6 has a new jobs system and like Kovid, I use Linux. What you might be seeing with adding might be a Windows specific bug in the new code.

Yeah, I was converting to MOBI, and yes this is under Windows XP. After converting a couple of PDFs in a batch Calibre would quickly shoot up to 2 GB in size and bring the system to a crawl. I've also just sit there and not do anything when adding a lot of new files at once, but I've seen this a couple of times with other types of files as well. Perhaps this is Windows-specific or related to drag-n-drop in the GUI.

[kovidgoyal]

Hmm windows issues, something I've come to know and love I'll look at it after fixing the various outstanding conversion bugs, since I don't think fixing it will be simple.

[sirbruce]

Kovid,

Thanks for all the fixes made so far; can we get a new beta build tonight or tomorrow?

[kovidgoyal]

Yeah just a couple more things to do before putting out the next build

[sirbruce]

Exxxxcellent.... yes, mmmmm, yessss....